Helmet Development and Standards

Edward B. Becker Executive Director Snell Mem orial Foundation 3628 Mad ison Avenue, Suite 11 North H ighlands, CA 95660 Tel 919-331-5073 FAX 919-331-0359 email: ed@smf. org

An Excerpt From:

Frontiers In Head And Neck Trauma Clinical and Biomedical

N. Yoganandan et al. (Eds.) IOS Press, OHMSHA
1998

Becker

Table of Contents
Summ ary Introduction The M odern Age A Brief H istory.. . Legal Influences Helm et Standards - Tests Helmet Impact Testing - Input Impact Te st Output and Evaluation Impact Test Apparatus Headforms Impact Surfaces Impact V elocity Oth er H elmet T ests. . Helmet E ffectiven ess Bibliography 1 1 2 3 10 11 11 14 17 20 22 23 23 25 27

Becker Summ ary The application of medical and engineering principles to the development and evaluation of traum a protective helm ets is trac ed fr om the 1940' s to the pr esent. The developm ent of per form ance standar ds, standard s organizations, standards program s and the relation of all three to commer cially available helmets is discussed. Par ticular attention is addr essed to the developm ent and centr al issues of methods for testing helmet per for ma nce in im pact. Introduction The basis for any pr otective device is two-fold: there m ust be the perception of risk and there m ust also be the perception that the device somehow attenuates that risk. From a ntiquity to the present, protective headg ear prevail whenever both these perceptions are present and disappear whenever either per ception is questioned. This basis is particularly true for tra uma pr otective headgear. M ilitary headgear provide an excellent illustration. The risk of head injury, particularly in warfar e, has been acknowledged throughout his tory. The use of protective headgear may be as old as warfare itself. Gurdjian [1] recounts head injuries mentioned in the Iliad and refers to accounts that Alexander the Great had been saved m any tim es by his fluted helmet. He describes military helme ts in use thirteen centur ies BC. and trac es developm ents through to modern tim es. How ever, protective helmets almost disappeared from combat after the rise of the musket. H eadgear that had proven useful against swords, slings and arr ows, served little purpose dealing with thrusting weapons like the bayonet or flat trajectory missiles like musket rounds. By the wars of the American Revolution, helmets had either disappeared or had shed their protective functions to serve other purposes. T he armies of the United States Civil War were issued cloth hats and cap s throughout the conflict. The re-emer gence of the metal combat helmet at the end of the nineteenth century may have been due to the advances medical treatment had made against bullet and shr apnel wounds to the extr em ities and tor so. These helmets may have also been useful bump caps in the ramshackle constructions built into the trenches of Wor ld War I. However, they probably owed their existence to the new economy with which industry could produce them. They wer e inexpensive, r ugged and identical; ideal issue for the millions that would be fielded in this century.

Becker The protective capabilities of all headgear inc luding com bat helm ets ar e continually being balanced by their wearers against other features such as visual impact, comfort and ease of use. The impor tance of visual impact, what sociologists might describe as the headgear' s ceremonial and decor ative function, is particular ly strong in our species. Although the head contains a cluster of sensory organs and the brain itself, all of which should merit protection, it also contains an elaborate signalling device, the human face. Our eyes are dr awn to others' heads and faces for identification. Facial r ecognition eve n has a unique locus in the bra in separate from other functions. T he visual impact of headgear w ill always weigh heavily in subjective evaluations of its worth. The M odern Age Wester n civilization is now just over fifty years into a new age of protective helmets. Pr otective headgear are comm onplace for activities for which bare heads or cloth caps were once considered sufficient. This helmet renaissance owe s much to technica l advances in trauma care and materials science but these are not the driving force. The source of this renaissance is that helmets have finally come to the attention of the same analytical spirit that revolutionized Western science and industry. Epidem iology is now providing strong objective evidence to support the two perceptions so basic to protective helme ts: that injury risks exist and that helm ets are effective counter measure s. [2, 3, 4, 5, 6] Me dicine and engineering are uniting to im prove he lme t protection. Governm ent and private organizations are fostering the development and sale of effective helmets to consumers. Helmet evangelists are preaching to the multitudes and lobbying legislatures and gover nme nt agencies. Although protective helmets have been used to advantage for more than thre e m illennia, the fir st system atic investigations of helmet function and effectiveness appeared only recently, in E ngland in the 1940' s. Ca irns in 1941 repor ted that in a study of over a 100 motorc yclist fatalities, 92% suffered from head injury and 66% had m ultiple injuries [7]. H e also discussed 7 cases of nonfatal injury in which helmets had been worn and in whic h th e in ju ry h ad b een " un usu ally mild." He discussed the structure of the helmets, noted accident damage and speculated as to how the helmets may have intervened to pr event mor e serious injury.

Becker Even in the 1940' s, motorcycle crash helmets had been available for some time. The helmets worn by Dr. Cair n' s subjects were described as crash helm ets, w ere British Arm y issue and differed substantially from standard combat gear. Dr . Cairns did not discover the crash helmet but he demonstrated conclusively that motorcyclists were exposed to a substantial risk of ser ious hea d injur y and that cr ash helmets could be used to attenuate this r isk. He a lso began the process of relating the mechanical behavior of crash helmets to the mechanism s of head and brain injury. Before Dr . Cair ns, helmet effectiveness w as anecdotal and helmet design was based on intuition and arm orers' lore. His 1941 [7] and 1943 [8] papers established the value of crash helmets as head protection and declared them fit subjects for m edical and engineering study. A B rief H istory of H elmet S tandards and P rogram s After World W ar II the M inistry of Transport in G reat Britain began a ser ious effort to investigate cr ash helme ts. The Road R esearch L aboratory of the D epartm ent of Scien tific and Industrial Research sear ched the

scientific literature for inform ation on head injury mechanisms and the mechanical properties of human tissues. They also conducted series of experime nts to identify potential helmet materials and helmet test methods. T heir work led directly to the first perfor mance standar ds for pr otective helmets. The first of these standards was British Standard 1869: 1952, Cr ash Helm ets for Racing M otor C yclists. [9] It was followed by British Standard 2001: 1953, Pr otective Helmets for M otor C yclists; British Standard 2095:1954, Industrial Safety Helmets (Light Duty); British Standard 2495:1954, Protective Helmets and Peaks for Racing Car Dr ivers and British Standard 2826: 1957, Industrial Safety Helm ets (Heavy D uty) [10,11, 12, 13]. Unlike earlier specifications which de fined objects in term s of their m aterials, dimensions and production, perfor mance standar ds defined helmets largely in term s of their function. That is, instead of describing the helmet they told how to test them. This per for ma nce testing was a new concept. He lmets were being presented as an intervention in a chain of accident dynam ics that would other wise lead direc tly to injury. Break the chain and pr event the injury. The test methods did not simulate the entire accident but instead attempted to reproduce the significant dynamics at the instant

Becker before the helmet intervened. The test outcome was then based on measur es of the significant dynam ics just after the helmet intervention. The nature of the intervention itself suggested appropriate test inputs and outputs. The tests for motorcycle helm ets applied shock loadings to a helmeted headform . The test technicians would drop a hardw ood block weighing ten lbs. fr om a height of nine feet onto a helmeted headform . The output consisted of dyna mic force m easur em ents recorded from a gauge mounted between the base of the headform and a massive reaction block. The test criterion required that the output for ce not exceed 5000 lbs. These standards served two imm ediate purposes: they wer e tools for the evaluation of available headgear and they also served as guides for the design of new headgear. The stresses of the accident an d estim ates of human tolerances had been translated to engineering term s directly applicable to helmets. H owever , the standards were and remain elements in efforts to regulate the manufacture, marketing and use of protective headgear. Qualifying helm ets wer e to be mar ked to identify the manufacturer , country of origin, helmet size and the number of the British standard. The helm et was also to bear the kite shaped certification mar k of the British Standards Institution (BSI). This ' kite mark' could only be used under license obtained from the BSI and required the

manufactur er par ticipate in quality assurance and testing program s administered by the BSI. Certain products must bear the BSI certification m ark to be sold at all in England. Ce rtain activities, such as motorc ycling, r equire the use of equipment bearing the mark and certified to the appropriate BSI standard. Even when the m ark wa s not legally requir ed either for sale or for use, the BSI ' kite mark' served as a guide to English consumers, distributors and retailers concerned about the capabilities of protective headgear. In the United States, helmet development was pursued m ostly by the military. By the late 1940' s, the U. S. Navy was funding investigations into head impact at at least two universities. One of these investigations was conducted by Dr . C. F. Lom bard, who or iginated the use of expanded polystyrene as a helmet m aterial. He and his colleagues at the U niver sity of Southern California studied shocks applied directly to the helmeted heads of research personnel and graduate students [14].

Becker However, these US efforts did not lead directly to the development of headgear for civilian use or to perform ance standards like those being formulated in England. The impe tus for this effort arose as a r esult of the death of W illiam Snell in 1956 in an amateur auto racing accident in northern California. Snell died of head injuries sustained in what was descr ibed as a survivable accident. His crash helm et had failed to protect him [15]. At the urging of mem bers and officials of the Sports Car C lub of Am erica (SCCA ), George Snively began an investigation of crash helmet perform ance. Snively, a m edical doctor, was an SC CA Cou rse Physician and had already been investigating protective headgear on his own for som e two years. He began a sur vey of auto racing headgear that precipitated a revolution in the helmet industry. A magazine article, " Skull Busting for Safety", appeared in the July 1957 issue of Sports Cars Illustrated detailing Snively' s findings. Snively found fault with almost every auto r acing helmet then available but, r emar kably, the crash helmet industry was receptive to the criticism. In a note published with the article, the magaz ine' s editor observed that the helmet industry had almost unanim ously gone into emergency operation to impr ove crash helmet perform ance. That same year , the Snell M emor ial Foundation was incorporated as a non-profit org anization in or der to sponsor Snively' s continuing wor k in crash helmets. By 1959, the Foundation had published the first Amer ican perform ance standard for pr otective helm ets [16]. In the early 1960' s the Foundation began to administer a helmet certification progr am sim ilar in some ways to the progr ams of the British Standards Institute but with a fundamental difference. The British Standards Institute set perform ance levels that every crash helmet should satisfy. Snively and the Snell M em orial F ounda tion set higher levels that only the best helmets would m eet. As mor e and more helm ets began to meet the standard, Snively would revise the test levels upwar d. He intended to create a continuing revolution and hurry the industry toward the best helmet a driver could be expected to wear. The motivation for manufacturer s to participate was purely economic. Snell certification helped sell helme ts. Snively and the Foundation adapted free market principles to promote the rapid evolution of headgear.

Becker Thr oughout the 1960' s the Foundation' s helmet standard was r evised steadily upward. Dur ing this time, the standard was also take n up for motorcycle helmets. Fr om 1970 to the present, the revision cycle stabilized to five year periods. As part of the 1985 revision, the standard was split into an ' M' series for motorcycle helmets and ' SA' for special applications which applied to auto racing helmets [17, 18, 19, 20, 21, 22, 23, 25, 26]. Other developments followed. In 1961 the American Standards Association (ASA) established a committee for protective headgear [27]. ASA, later the Amer ican National Standards Institute (ANSI) was and remains an um brella organization which prom otes standards development for a broad range of products, ser vices and activities. AN SI polices the form ation of committees and the standards writing process. They ensure that the committees are balanced over all facets of interest including providers, consumers and knowledgeable individuals. They require standards to be written according to a given format and subject them to a general ballot before adoption. Once a standard is adopted, AN SI requires that the committee reconsider the standard at least every ten years m aking revisions as necessary. AN SI then conducts the same genera l ballot before the standard is readopted. AN SI Standards are consensus standar ds. Financial interest is not a barrier to participation. Some of the most energetic par ticipants in the process are the manufacturer s of the very products to which the ANSI standards apply. As a r esult, AN SI requirem ents should be met by ever y product and service. How ever, AN SI does not adm inister any cor respo nding cer tification progr ams. The two standards makers discussed previously, the British Standards Institute and the Snell Mem orial Foundation, each conduct pr ogr am s involving pre-m arket and follow-on testing for all those products for which certification is claimed. Before a manufactur er can adver tise Snell certification or apply the BSI kitemar k, he must submit to the standards policing pr ogra ms that eac h of those organizations adm inisters. AN SI imposes no such obligation. The manufacturer himself determines whe ther he m ay claim qualification to ANSI helmet standards. He need not provide any supporting documentation or even notify ANSI of his claim. The first ASA helm et standard was Z 90. 1-1966, Pr otective Headgear for Vehicular U sers published in 1966 [28]. The first revision w as published by AN SI in 1971. A supplem ent, AN SI Z90. 1a-1973 was released in 1973 in order to cor rect a technical error [29].

Becker The International Standards Organization (ISO) also for me d a technical subcom mittee to consider protective headgear in 1960. T his activity led directly to the prom ulgation of ISO Recom mendation R 1511, Pr otective He lmets for Road U sers in 1970 [30] and, later, of ISO draft standard, DIS 6220-1983, Headfor ms for use in the testing of protective helmets [31]. ISO itself arose in 1946 out of a United Nations effort. ISO standards are intended to promote international trade and ar e recom me nded as m odels for gover nm ents and other s to use in de veloping national standards. Like AN SI, ISO publishes standards but does not adm inister any cor respo nding cer tification progr ams. Another United Nations effort produced R egulation No. 22, Uniform P rovisions Concerning the Approval of Pr otective He lme ts and of Their V isors for D rivers and P assengers of M otor C ycles and Mopeds [32]. This was part of a general agreement concerning motor vehicle equipment and parts enacted in 1958. This docum ent is

cur rently in its fourth revision and, like ISO standards, is intended to prom ote international trade by harm onizing standards and enabling mutual recognition of appr oval. In 1972, the United States Government announced a draft motorcycle helmet standard, Federal Motor Vehicle Safety Standard 218 (FMVSS 218) which would come to be known as the DO T standard [33]. The draft was taken almost directly from the m ost recent revision of the ANSI standard, Z90. 1-1971, but included plans to impose more string ent r equir em ents in Septem ber , 1974, eighteen months after the standard wo uld take effect. While the DOT draft was still being considered, AN SI published the supplement, Z 90.1-1973a, to repair a technical flaw discovered in AN SI Z 90. 1-1971. The 1971 standar d specified a newer test method but applied c riter ia developed for an older method. AN SI Z90. 1-1971 was more difficult as written than the comm ittee had intended. Further mor e, man y consider ed that the additional difficulty would not necessarily lead to better helmets. The Na tional H ighw ay T raffic Safety Adm inistration (N HTSA), the responsible agency w ith in D OT announced in 1973 that they would continue with the original criteria from AN SI Z90. 1-1971 but did defer any decision on the changes scheduled for 1974[34]. The D OT standard took effect in 1974 essentially unchanged from the original draft except that the scheduled changes had been dropped altogether [35]. Although slight revisions have

Becker been ma de since its inception, the standard rem ains essentially unchanged from its original form. How ever, NH TSA has recently begun to consider an overhaul [36]. Like the BSI m otor cycle helmet standard, the DOT standard is mandatory. Once it took effect, every helmet sold for use in street motorcycling in the United States was required to meet it. However, like ANSI, manufacturers claimed the certification for themselves. Once the manufacturer had claimed DOT qualification for his products, he was obliged only to label them with the DOT emblem. M anufacturers w ere not requir ed to make any submission of samples, test data, production recor ds or even notify the government before introducing a new helm et onto the ma rket.

A group of motorcycle helmet manufacturers formed an industry organization, the Safety Helmet Council of Am erica (SH CA ), to provide a third party certification program to the new DOT standard. The program required manufacturers to subm it test data for each new motorc ycle helmet model before introducing into the m arket and to make annual submissions of test results for each model already on the market. Participating manufacturers were allowed to m ark their pr oducts with an SH CA label. The SH CA collapsed dur ing the 1980' s. When the DOT standard took effect, it was accompanied by a Feder al push for State laws re quiring motor cyclists to use appropriate headgear. M andatory use laws wer e enacted in a number of states but many were repealed a few years later. For a time, NH TSA linked distribution of Federal highway monies to the passage of state helmet laws. This practice has been disallowed. Bicycle helmets wer e also attracting attention. In 1970 BSI published British Standard 4544:1970, Pr otective Helmets for P edal Cyclists [37]. In 1972, the Snell Mem orial Foundation released three appendices to their 1970 general helmet standard, one of which applied to bicycle helmets [38]. The Snell standar d wa s revised upward in 1984, 1990 and 1995 [39, 40, 41]. In 1984 AN SI published a bicycle helmet standard, Z 90. 4-1984 [42]. In 1993 the Am erican Society for Testing and Materials (ASTM ) published a bicycle helmet standard, F1447-1993, w hich was revised a year later in F 1447-1994 [43]. C urr ently, the Consum er P roduct Safety Com mission (CP SC), an agency of the U S F eder al governme nt, is dra fting a bicycle helmet standard which may take effect in the spring or summer of 1998.

Becker ASTM is an um brella standards organization very similar in scope and practices to ANSI. As with AN SI, ASTM standar ds ar e consensus standar ds that every pr oduct should be expec ted to m eet. Also, as with AN SI,

manufacturers determine for themselves whether they may claim ASTM qualification and then m ay proceed to do so with no subm ission of docum ents or notification of AST M authorities. CP SC began drafting a bicycle helm et standard as a r esult of an act of C ongr ess, The C hildren' s Bicycle Safety Helmet Act of 1994. The C omm ission had been petitioned directly to do so in 1989 [44] but had rejected the petition. A re view of the US bicycle helmet industry and of existing voluntary standards and programs for bicycle helmets had persuaded C PSC that regulatory action was not justified. CP SC has circulated two successive draft standards since the act was passed. T he helmet requirem ents seem well reasoned and are stated clearly. As with FM VSS 218, the DO T m otorcycle helmet standard, manufacturers will determ ine whether their pr oduc ts qualify and will then proceed to claim the qualification. CP SC r equires no

submissions or notifications but does oblige m anufactur ers to m aintain a set of test recor ds to support their claim s. Football helmets came in for scr utiny particularly after r ising trends in head and neck injuries were observed in the 1960' s. In 1973, two medical doctors, H. A. Fenner and A. F. James published a football helmet standard, JF73 [45]. The for ewor d to the standard states that it had been prepared and pr inted at the personal expense of the authors. The authors had previously participated on an AN SI com mittee established for the specific purpose of prom ulgating a football helmet standard but the committee had been disbanded just when a final draft seemed near. JF73 was and rem ains a particularly demanding standard but it was well within the technology that existed at the time.

In 1975 the National Operating Com mittee for Sports and Athletics Equipment (NO CSA E) published a football helmet standard that has since gained wide acceptance in the United States. M ore recently, the Amer ican Society for Testing and Mater ials has also published a football helmet standard [46]. Com peting standards have proliferated in the United States and throughout Europe. In the US the effect is confusion. Since all the standards and program s promise safety, manufacturers and users often apply other criteria.

Becker Manufacturers choose whichever m akes the most business sense while user groups often select for helmet aesthetics.

In Eur ope the standards acted as trade bar riers. One of the changes the Comm on European M arket imposed on its mem bers was the form ulation and adoption of new Com mon Eur opean Nor ms or CE N Standar ds. In so doing, the Com mon M arket will enforce mutual recognition of standards, certifications and products among its member nations. These C EN standards are to be taken from ISO Standards whenever possible. The system is administered by ' notified bodies' that is, agencies throughout Europe emp owered to consider applications for acceptance and to award the ' CE' mark which identifies each product mee ting appropriate requirem ents. O nce a product is CE mar ked, it may be tr anspor ted and sold freely throughout m ember countries. For protective equipment and, particularly, helm ets, there is still much con cer n that the C EN standar ds will repr esent an am algam of the least stringent national standar ds. Although one or more ' notified bodies' have been designated in each mem ber country, there is much uneasiness that manufacturers and others may dir ect applications to the least demanding of them. Finally, there is no provision for proving the performance of products already awar ded the CE mar k. Legal Influences There is another uniquely Amer ican aspect of helmets that is also catching on in Eur ope, helmet liability. Helmet man ufactur ers, distributors and re tailers m ay be liable for damages if a helmet fails to protect its wearer. Since the consequences of head injury are often severe, damage awards can be very large. T he industry has turned to insurers to provide liability coverage. They have tur ned to standar ds as a means demonstrating the due diligence necessary to produce effective headgear which may in turn obtain lower insurance premiums and favorable court judgem ents. When a more stringent standard exists, it may not be sufficient to produce helmets to a lower standard even if that standard is set by the government. Although there have been attem pts to hold standar ds m aker s responsible

10

Becker for inadequate standards, the manufacturer , the distributor and the retailer have usually been held responsible to select and adhere to a proper standard. Evidence presented by the plaintiff in helmet liability cases often includes performance testing conducted on samples of the model. Fr equently, such testing is the first manufacturer -independent testing perform ed. Except for the cer tification pr ogr am s of the Snell M em or ial F oundation and those of the Safety Equipment Institute, all the curr ent United States product qualifications are claimed by the manufactur er. NH TSA has done some spot checking of the performance of DOT labelled motorcycle helmets in the US mar kets but far too little to be considered effective policing. Thus, civil liability is often the only check on lar ge segme nts of the helmet industr y. Since the issue in civil liability is whe ther and in w hat am ount dama ges ar e due, it may not b e a r eliable means of rem oving ineffective headgear fr om the mar ket or en cour aging the industry to produce better, mor e protective helmets. Helmet Standards - Tests Standards facilitate trade. T hey codify expectation for both provider and consumer and relate the expectation to measurable product attributes. Standards require either that there be some gener al agreem ent on expectation and attributes or confidence in the integrity and capability of the standards maker. Since the BSI standards issued in the early 1950' s, every helm et standard specifies tests for protective perform ance. Although each standard specifies tests for several different aspects of helmet performance and some standards test aspects not considered by any other standar d, all the standards specify tests for impact protection. Helmet Impact Testing - Impact Input Impact protection is the prim ary consideration of almost every helmet standard. The prescribed tests seek to reproduce the significant aspects of impacts to a helmeted head and then to measur e and evaluate the significant aspects of the outcome. The impacts as reproduced are highly simplified interactions. V irtually every test is a one dimensional exchange of mom entum between a helm eted headform and an impactor. Rotations are minimized by careful alignment of the centers of gravity of the impacting bodies, careful alignment of the impact surfaces and sometimes by the

11

Becker mechanical test apparatus itself. It is generally considered that this one dimensional test configuration is sufficient to evaluate impact protective performance but one standard does prescribe an oblique impact test which is discussed fur ther along in this developm ent. The standards also specify the masses of the impacting bodies, the headform proper ties, the surface configuration and properties of the impactor, and wher e on its surface the helmet m ay be impacted. Finally, the standards call out impact severity. The critical differences between standards in the issues discussed so far involve the impact sites, the impa ctor surfaces and the imp act sever ities. Impact sites are m ost often prescribed by procedur es that map a test area onto the surface of the helm et. The helm et is first placed on a special headfor m a ppr opr iate to the helmet size. P lanes and points in the headfor m ge ome try ar e trace d on the helm et surface. These mar kings then guide the construction of a test line. Test impacts must be sited on or above the test line. Thus, the headform determ ines the impact sites rather than the helm et. Since the helmet may extend below the test line, especially on the sides and back , ther e is some concern that users may infer pr otection from lower im pacts. H owever , most standards maker s resist the conclusion that helm ets should be tested over their entire surfac e. Ther e are two r easons. The fir st is that although im pacts involve ar eas, their sites are descr ibed as single points near the center of the impac t area. An im pact on the test line must

necessar ily include an area that extends below the test line. Whether stated or not, this impact ' footpr int' is usually a consideration in the test line definition. The second r eason is that parts of the helmet extending below the test line almost always provide some protective benefit even if no t at the level req uired on and above the test line. If these parts are not specifically required but will be tested if present, a manufacturer could conceivably tur n a failing helm et into a passing one me rely by trimming these extensions away. A standard that rejects more protective headg ear in favor of less pr otective units is not acceptable. Alm ost every standard specifies a flat impact surface for some of the test impacts. M ost also r equir e impacts against a shaped surface that delivers a much more focused shock to the helmet surface. Traditionally, this shaped

12

Becker surface has been a spherical section with a radius of curvature of about 48 mm how ever, other shapes are also used either in addition or as an alternative. T hese shaped surfaces challenge the helm et in ways that the flat anvil does not. The use of both varieties may also facilitate better helmet evaluation given the nature of the instr um entation used to monitor the test response. Impact severity is generally specified as a velocity mea sured just before the im pact. M ost descriptions usually include a theor etical dr op height. However , ma ss is also an im por tant par am eter in the specification. T he Snell M em orial F ounda tion typically specifies impact severity in term s of energy but this ener gy r educes to a velocity requirem ent once the impact mass is specified. Impact severity requirements vary widely across helmet types and even across standards prepared for the same helm et type. At this tim e, ther e ar e no objective determ inations for prescribing impact severities for helmet tests. Although some suppose that certain activities may actually require less protection than others, even the m ost demanding standards state that helmets meeting all the requirem ents may still fail to pre vent injur y or death in reasonably foreseeable accidents. H elmet standards do not specify all the protection a person might reasonably need. At best, standards specify only as much helmet as a person might reasonably be expected to wear. If ideas such as style and trad ition ar e not consider ed, reasonable expectation might lead standards makers to consider much higher requirem ents for impact severity. However , since he lmet thickness necessar ily incr eases with test sever ity, people will often refuse to wear serviceable helm ets in favo r of less pr otective but m ore aesthetically pleasing headgear. There is a tension between these two ideas: all the helmet people can wear versus all the helmet people will wear. If standards cannot resolve the issue they should at least maintain the tension. Fashion is malleable. It will yield, how ever slowly. Impact Te st Output and Evaluation

13

Becker The prec eding account describes most of the issues in impact test inputs. Ther e are also impact test outputs. Gener ally the outputs consist of a single axis of force or translational acceleration for the entire test headform . These outputs are captured and compared to a test criterion in order to evaluate the helmet' s performance. There are several criteria in use. The two simplest ar e peak cr iteria and du ra tion cr iteria. Peak cr iteria require that the highest value of force or acceleration recorded does not exceed so m e m a xim um perm itted value. Dur ation criteria limit the amount of time for which the output may exceed a specified value. Some standards use m ore com plex criteria such as the Gadd Severity Index (GSI)[47] and the Head Injury Cr iterion (HIC)[48] w hich ar e em pirical attem pts to relate injury to the time history of head translational acceleration. There has always been uncer tainty w hether these cr iteria are any more r eliable than peak criteria or whether they even apply to helm eted he ad im pacts at all. H owever , the biggest barr ier to their use had been computational complexity. Now that computers are used r outinely to acquire test outputs, HIC is receiving more attention. All these cr iteria attem pt to relate the test output to injury. How ever, there is little solid information on which to base these cr iteria. The peak cr iteria seem to have come about through exper iments producing skull fracture in cadaveric heads and Geor ge Snively' s investigations of auto racing accidents. The first BSI standards stated that the human head could withstand forces on the order of 5000 lbs and set peak force levels accordingly. Th is level corr esponds to accelerations of about 500 times gravity (500 G' s) which has been revised downw ard in succeeding standards to 300 G' s. The time duration criteria m ay have been an attem pt to acknow ledge infor ma tion developed at W ayne State Univer sity in the 1960' s showing that the tolerance of the human head to force var y with the duration of the exposure. The first time dur ation criteria were w ritten into the ASA Z 90. 1-1966 standard for auto racing he lm ets [28]. However, these time dur ations did not have any practical effect sinc e none of the helmets of the tim e ever failed to meet them. The 400 G peak was the only significant criterion in ASA Z 90. 1-1966. The succeeding ANSI Z 90. 1-1971 standard [29] introduced changes in the test method which increased the values of the durations observed for all helmets. The dur ation criteria suddenly began to eliminate many of the

14

Becker products that had easily met previous requirements. Since this effect was entirely unexpected, the ANSI comm ittee released AN SI Z 90. 1-1973a w ith new values for the dura tion criter ia adjusted to match the new pr ocedur es. The Gadd Sever ity Index was an attempt to reduce injury data from a number of sources to a single algorithm that could be applied to time histories of head acceleration data. T he essence of the calculation raises acceleration in G' s to a power of 2. 5 and integrates with respect to time in seconds. So long as the product of this operation does not exceed 1000, the acceleration pulse was considered noninjurious. GSI was used to set design and eva luate vehicle interiors wher e bare headed im pacts with dashboards and other obstruction had been a growing concern. The Head Injury Cr iterion is a refinement of GSI that was adopted as Federal Motor Vehicle Safety Standard 208 (FM VSS 208). The calculation now raises average acceler ation to the 2.5 pow er and m ultiplies by the length of the time period in seconds. T he peak value for any time period of up to 32 milliseconds over the duration of the acceleration pulse must not exceed 1000. The first drafts of the DOT motorc ycle helmet standard FM VSS 218 intended eventually to substitute HIC for peak a nd dur ation cr iteria taken from the A NSI Z 90. 1-1971. However , HIC would have eliminated every helmet then available. When the standar d took effect, HIC was dr opped but the author ities also refused to revise the duration criteria to those of ANSI Z 90. 1-1973a. Recently, United Nations Regulation 22 Revision 4 [32] applies the H IC algor ithm to evaluate m otor cycle helme ts. The m aximum HIC value perm itted is 2475. This level is well above the 1000 limit set in FM VSS 208 and is also well above the level of 1500 advanced by so me as appropriate for helmet tests. It does appear to be w ithin the capabilities of many currently available helmets. Even when there is agr eem ent on the type of cr iterion, standard s differ on the criter ion limits. Some of the differences reflect concerns about the needs of the populations who will wear the helm ets. Som e ar e in response to injuries like slight concussion which had previously not been considered threats to life or quality of life. F inally, some believe that helmets may be evaluated reliably at substandard levels of impact severity mer ely by applying more stringent criteria.

15

Becker M ost helmet criteria are based on the needs of adult males. The inform ation available is sparse but virtually all of it com es fr om accide nts involving young adult ma les and tests conducted on adult m ale subjects or adult m ale cadaver ic segments. Older people may have different needs. It has been noted that aging tissues becom e less flexible and that older pe ople require longer r ecover y times for similar injuries. Ver y young children may also have different needs but, for these, the picture is not so clear. Children' s tissues are m uch mor e flexible than those of even young adults so that higher levels of force and acceleration may actually be tolerable. Helmet criteria are also based on levels of force or acceleration thought to produce death or pr ofound nonrecoverable injury. Concussion had only been considered when the circumstances of the incident required that the wearer be able to execute escape ma neuver s. Com bat pilots, for example, need to be able to activate ejection seats, exit sinking or burning aircr aft and possibly evade hostile forces afterwar d. However, multiple concu ssions and even multiple blow s of subconcussive str ength may lead to perm anent injuries [49]. For this reason, some have suggested that helmets should be required to pr event concussion. These same people have also suggested that concussion protection requires softer helmets than those currently available. Therefore, test criteria must be made more stringent so as to force the production of softer headgear. If softer helm ets ar e to withstand test impacts at current severities, they must be substantially thicker. Otherwise, som e of pr otection from the sor ts of head injuries that imm ediately kill or disable m ust be sacrificed. It is unlikely that the public would immediately accept thicker helmets. Currently, there is no reliable information concerning test criteria and the risk of concussion. Ther e is no basis either to set new criteria to reduce the incidence of concussion or to evaluate a trade off between concussion protection and the sorts of protection for which helm ets traditionally have been worn. Finally, som e suppose that im pact sever ity and impact criter ia are som ehow linked so that tests conducted at lower levels of impact may still be made useful merely by evaluating the results according to more stringent criteria. There is no such linkage. Helm ets may test well up to a certain severity but beyond that the outputs spike upward beyond any test criteria and, usually beyond the range of the instrum entation. The transition is sudden. It depends

16

Becker on the test severity and on the design and construction of the helmet. Ther e is no known way to examine output for a helmet tested at a severity below this transition and determine per form ance at higher levels of severity. Impact Test Apparatus There have been many different systems for per form ing helmet impact tests since the 1940' s. The following descr iptions detail the progr ession to curr ent methods. The Brinell test [11, 13] employs an apparatus that resembles an oversized office stapler. A headfor m is supported atop the stapler arm so that shocks applied to the top of the headform press a hardened steel ball into the surface of a small aluminum impression bar in much the same way that a staple is pressed into sheets of paper. A helmet is placed on the headform and an iron ball is dropped from a measured height onto its crown. The diameter of the im pression left in the alum inum bar is directly related to the peak for ce tr ansm itted thr ough the helm et. There are several disadvantages. T he most serious is that only crown impacts are well accommodated. Cair ns and almost every investigator since has pointed out that cr own impacts are rare. Cr ash helmets receive most impacts in the front and, to a lesser degree, on the sides and rear. This Brinell technique was used well into the 1950' s in the BSI standards for industrial safety helm ets. Ho wev er , the BSI standards of that time for motorcycle and auto racing helmets [10,11, 12] made sever al improvem ents on the technique. T he iron ball was replaced with a heavy block that was guided to impact by w ires. The stapler mechanism was discarded. Instead, the headform w as mounted directly on a force transducer that converted the vertical com ponent of for ce into an electric signal that could be captured and analyzed. The headform itself had been redesigned to perm it helmet impacts in the brow and rear as well as the crown. Essentially, the helmet could be positioned w ith either the br ow, the cr own or the rear upper mo st. These improve ments corr ected many of the problems associated with the Brinell device. The lower surface of the heavy block could be shaped to sim ulate a ran ge of imp act surfaces. The entire time history of the force was available for study. However, the impac t sites on the helmet were still limited. H at band impacts, that is, lateral impacts and impacts low on the brow or the rear of the helmet were still not possible.

17

Becker After experimenting with pendulum devices, G eorge Snively at the Snell Foundation began to work w ith a ' swing away' test rig that would facilitate these hatband impact configurations. The or igin of the swing away r ig is uncertain. Corr espondence from the m id 1970' s [50] suggests tha t it was developed by the Snell M emor ial Foundation but neithe r Sn ively nor the Founda tion has ever claim ed cr edit. The swing away device replaced the floor mounted headform and force transducer with a headform mounted on a pivoting armature. The arm ature, stabilized in a horizontal orientation by a brittle glass rod, held the headform in position until it was struck by the wire guided impa ctor. At the instant of impact, the rod shattered and allowed the arm to swing down and away under the force of the blow. An accelerom eter m ounted at the center of the

headform produced an electronic signal proportional to the headform acceleration and to the forces applied to the headform. The advantage was that impacts could be delivered easily to the front rear and sides of the headform. The disadvantage was that the swing away device was m echanically com plex and many of its param eter s directly influenced the test. The interaction was between two movable inertial bodies and differed significantly from the one body systems used before and since. Although the system was technically feasible, it did not have the intuitive appeal of previous methods or of the ones that followed. Com parisons betwee n swing-aw ay tests an d tests on other device s have been plagued with misunderstandings mer ely because many people failed to consider the mechanics of the intera ctions. One of these misunderstandings led ultimately to the time duration controv ersy in the A NSI Z 90. 1-1971 and F M VSS 218 standar ds. Snively and the Foundation moved on to falling headform devices in the mid-sixties but swing away devices were used in England through the Seventies. There are now a number of falling headform devices being used in helmet impact testing. A helmet is tested by placing it on a headform of a given mass and allowing it to fall onto an appropr iately shaped anvil supported by a massive reaction block. As with the swing away devices, the headform response is taken from an accelerometer mounted at its center. The m ethods are appealing because of the obvious similarity with accidents in which a falling person' s head strikes a rigid unyielding surface.

18

Becker There are two br oad classes of falling headform devices, guided fall devices which control the orientation and position of the headform and free fall devices in which much of that control is foregone. The advantage of the guided fall devices is that a single axis accelerometer may be used. A cceleration is a vector quantity with two horizontal components in addition to the vertical component. Since the vertical component far outweighs the other two, guided fall systems are set to orient the accelerometer to capture this component and the two hor izontal components are left unmeasured. There are two types of guided fall system s in use, twin w ire and m onor ail. In both, a metal ball is mounted on a fram e with bearings that slip along the wires or the monorail. The ball fits into a socket in the headform so that the headform m ay be a djusted in a br oad r ange of or ientations. A simple clamp m ay then be tightened to lock the headform in position. Since the orientation of the ball w ith re spect to the guidance system never changes, the

dynamics of any impact may be monitored by a single axis accelerom eter positioned inside the ball with the sensitive axis aligned along the direction of motion. One disadva ntage is that the helme t may inte rfere w ith the guidance mechanism for certain extrem es of headform position. Another is that the guidance device itself may com plicate the dynamics of the impact by

introducing extraneous resonances. Attempts to minimize interference problems usually add to the size and m ass of the guidance fr ame and incr easing the interference from these resonances. These problem s are par ticularly true for twin wire system s. The monorail also has some interference problems bu t the guidance fr ames are gener ally lighter and less resonant than for twin wire systems. Unfortunately, the bear ings are subjected to much gr eater stresses dur ing a test complicating maintenance and test repr oducibility. At this time, there is no clear choice among the various configurations of monorail and twin wir e systems. There are sev eral configurations of each type cur rently in use in the United States. Fr ee fall devices do away with guidance fram es and the ball and socket articulation on the headform. Instead, the helmeted headform rests over a hole on a platform. The entire platform is dropped in a guided fall, usually guided by three wires, toward the r igidly fixed anvil. The entire anvil fits through the hole in the platform and makes direct

19

Becker contact with the helmet. The helmet and headform are then free to m ove in r esponse to the impact while the platform continues to drop away. Since the dynamics of the event are over within ten to twenty m illiseconds, the helm et is often loosely held by a net or basket attached to the platform. The advantage is that there is no guidance frame or bearing system to interfer e with the positioning or with the impact dynamics. The disadvantage is that a full three axes of accelerometer data must be acquired because there is no r eliable m eans of cor rectly orienting a single axis transducer. Fur thermore, for any non-planar impact surface, it is almost impossible to position the center of gravity of the headform with respect to the surface axis of symm etry. The result is that many of the impacts on these surfaces are glancing blows that neither test the helmet to the maximum allowable lim it nor yield r eproducible results that could be com pared with other tests. Headforms The headforms themselves are a critical part of impact test systems. There are a num ber of specifications for impact test headfor ms. The two specifications comm only used in the U nited States are those in the DOT

mo torcycle helmet standard FM VSS 218 and ISO D IS 6220-1983. Ther e are three DO T headfor ms, small, medium and large. T he specification describes the external surface of the medium size and use s scaling factor s to generate the small and large sizes. The specified mass for each size is proportional to the cube of the scale factors. The source for the specification is uncertain but is rumor ed to have come from an anthropometric survey of US soldiers conducted in the 1940' s. The ISO DIS 6220-1983 specification includes separate descriptions of headforms starting at 50 cm circumference and increasing in circumference by one centimeter increments. These headforms are not geome trica lly similar; that is, they ar e not sca led fr om a single refer ence. Only four sizes fr om the range are com mo nly used but som e cur rent standards have added the ' O' headform, 62 cm circ um fer ence, to the standard set. The specification calls out a total mass of 5 kg regardless of headform size. However, CPSC and ASTM are considering lower m asses for headfor ms used to test children' s helmets. The cur rent C EN headform specification calls out headforms with ISO geometry but with masses proportional to the cube of headform circumference.

20

Becker Although both ISO and DOT headfor ms a re intended to correspond to W estern head anthropom etry, the descriptions are not at all similar. For headform s of the same circum ference, the AP (fr ont to back) length of the ISO headform is smaller, the breadth is greater, and the head height, the distance from the plane corr esponding to the anatomical Frankfort plane to the apex of the headform, is smaller. These headform issues of ma ss and geom etry ar e cr ucial. The m ass determines the total m ome ntum that m ust be exchanged in an impact involving a specified velocity differ ential. A qualified helmet may fail to m eet test cr iteria if the headform is too light or if it is too heavy. Hum an head m ass data collected by W alker [51] and Beier [52] does not support any correlation between head mass and head dimensions. What correlation there is seems to be between head m ass and whole body m ass. The cubic relationship, in spite of its appeal to engineering intuition, is clear ly not supp orted by anthr opom etric data. Headform geom etry is one of the determ inants of helmet coverage. All the impact standards define how impacts may be sited in terms of planes and points fixed in the headform . Slight variations in headform geom etry, misplacement of the helmet on the headform or the use of the wrong size headform will shift legitimate test impact sites and could conceivably cause adequate headgear to be rejected or inadequate headgear to be accepted. All the headfor ms just m entioned ar e to be made of a har d non-r esonant m aterial. The ISO headform requir em ents [31] originally specified wood but, curr ently, alm ost all the above are m ade of low resonance m agnesium or aluminum alloy. However , ther e is concern that rigid headform s may fail to duplicate the impact dynamics of the human head. Saczalski [53] has described a phenomenon he observed in computer simulations of helmeted head im pact. Dur ing simulated imp act, portions of the finite element head model bulged laterally outwar d perpendicular to the impact axis producing what Saczalski r efer red to as ' squashing' . Clearly, rigid headforms will not r epr oduce this response. How ever, whether squashing or som e other significant complex behavior actually takes place and actually bears on the injury outcome of a helmeted head impac t is, as yet, uncertain. What is certain is that compliant headforms are technically much mor e complex than their r igid counterp arts. There is no confidence that compliant test headforms can be adequately manufactured or even specified. Cur rently,

21

Becker only NO CSA E standar ds call for the use of a com pliant headfor m. The N OC SAE headfor m, which incor pora tes a soft outer surface and a glycerin filled brain cavity, has not gained general acceptance. In addition to these headforms, there are other headform specifications that address the needs of different ethnic groups. The Japanese standards [54, 55] , for exam ple, call out headfor ms with the greater breadth to length ratios than the headforms typically used in the US and Europe. Impact Surfaces The standards also specify a variety of impact surfaces. Impacts with flat surfaces are thought to be the most comm on accident configuration. W ith the development of the wire guided impactor, flat surfaces have been a part of every helm et test procedur e. The spher ical impactor used in the Brinell tests has been carried for war d to many curr ent standards as the hemispherical anvil. Other currently used surfaces include the curb stone, various cylinders, nar row edges, a shar pened right angle and a hor se shoe anvil. These anvils have bee n devise d to repr esent anticipated im pact hazards but the pairing of flat and hemispherical surfac es also serves another purpose. One of the protective be nefits of helmet use is that localized loadings applied to the helmet surface are distributed to a much wider area of the head. The impact tests described above cannot determine this effect directly because the forces and ac celer ations actually m easur ed app ly to the w hole headform. The tests yield the sum of the shocks applied to the headform rather than the shock applied to any specific area. It is conceivable that a helmet with no load spr eading capacity at all could be devised to meet the test

requir em ents for either the flat or the hem ispher ical anvil. However , one suc h a helm et could not m eet r equir em ents for both anvils. Although most helmet standards specify hard metal anvils, football helmets, hockey helmets and some others are frequently tested in impact against compliant surfaces. T he modular elastomer pad (M EP ) is composed of resilient but highly stable material usually molded into a pad of one inch thickness. M EP testing is considered nondestructive for many sports helmets intended to protect against repeated impacts. However, the MEP attenuates some the impact itself so that compar isons with tests against non-resilient surfaces may be m isleading. Impact V elocity

22

Becker In addition to drop m echanism s, headforms and anvils, most curr ent im pact test stands also incor por ate devices to measure the velocity of the falling headfor m just befor e impact. Current standards specify, in one form or another , the impac t velocities for each test configuration. D rop height is generally not specified because frictional losses before the impact make it an unreliable parameter for impact sever ity. A comm only used method to measure im pact velocity is timing the passage of a tab or two tabs through a light beam . A light sensing device is set to detec t the passage of the leading and trailing edges of the tabs. The velocity can be determ ined by measuring the time interval from leading to trailing edge or from the first leading edge to the second leading edg e and dividing into the measur ed distance betw een these edge s. Other Helmet Tests and Considerations In addition to impact tests, standards generally set labeling requirements for identification and w arnings. They also set some limitations on helmet configuration, call out environmental conditionings, set the number of samples to be tested and describe the uses for which qualified headgear may be appropr iate. Standar ds also prescribe tests for other kinds of perform ance. There are two types of retention system tests. Strength tests load the retention system components to ensure that they will not fail under certain levels of loading and, in some children' s helmet standards, to ensur e that the helmet will release to avoid hanging injuries. Positional stability tests attempt to determ ine whether the retention system will hold a helmet in position by applying a tangential shock load to snatch a helmet from a test headform. Standards also call out visual field requirements. The visual field is generally specified in term s of a solid angle referenced to a test headform. Auto racing helme t standar ds often call out flam mability test requir em ents. Comp onents of the helmet will be subjected to a flame and requir ed to self extinguish within a given time per iod. The shell penetration test m easur es a helmet' s ability to w ithstand a conical pointed im pactor . This test appear s in almost all mo torcyc le and auto r acing helm et standards from the earliest BSI specifications. The Australian and Swedish bicycle helmet standard s [56, 57] call out force distribution tests. The helmet is placed on a headform equipped with a force transducer so that only the force applied to a limited area of the

23

Becker headform' s surface will be measured. A shaped impactor of a given ma ss strikes the he lmet at a given velocity directly over the transducer. The peak load must not exceed a certain value. There are two appr oaches to testing facial pr otection provided by full face helm ets. Full face helm ets

incor por ate a chin bar, an extension of the helm et shell crossing laterally across the chin from the right to left temporal areas of the helmet. Snell Mem orial Foundation Standards apply a me asur ed shock loading to the chin bar of a r igidly supported helmet and measur e the maximum intrusion into the helmet interior. If the chin bar intrusion exceeds a certain limit the helmet is rejected. BS 6658:1985 P rotective He lme ts for Vehicle U ser s [58], which replaced separate BSI motor cycle and auto racing helmet standards, describes another chin bar test. BSI also applies a shock load to the chinbar but the helmet in this test is placed on a r igidly fixed full headform . The back of the helmet shell rests against a rigid surface topped by rubber pad. Thus the shock load ma y be tr ansm itted thr ough the helm et stru ctur e into the rubber pad or into the chin of the rigidly fixed headform . The deceler ation of the striker impar ting the shock load m ust not exceed 300 G' s. BSI standard BS 6658:1985 [58] also calls out an oblique impact test. The anvil surface is flat but slopes upward so that the falling helmet strikes glancing blow. T wo different anvil faces are used, one is a sheet of abrasive material securely clamped in place. The other is a series of parallel steel edges that convert the anvil face into a sort of cheese grater. Load cells in the anvil capture the forces parallel to the surface. The test criteria limit the peak value and the first time integral of this force. The standard states dir ectly that the test is intended to assess frictional forces and forc es resulting from projec tions. Glaister [59] describes the development of the test. H e states that it is intended to address concerns about angular acceleration. Angular acceleration had been identified as a potential injury mechanism since the 1940' s and is currently thought by some to be the primary injury threat. A lthough standard impact test perform ance implies a reduction in angular as well as translational acceleration, tangential forces applied to the helmet shell are not addr essed. In fact, the standard impact test configurations ar e gener ally chosen to m inimize tangential forces. However, although angular acceleration may represent an injury hazard, there is no indication that tangential forces are producing either high levels of head angular acceleration or head injury. Fur thermore, the oblique impact

24

Becker test as described cannot determine whether forces are transmitted to the headform and par ticularly what the resulting peak accelerations in the headform might be. Helmet E ffectiven ess Alm ost every critical assessment of crash helmet effectiveness has concluded that helmets reduce the risk of head injury. Hur t' s study [2] showed that motorcycle helmets pr otected their w earer s in a survey of accidents in Los Ange les. Williams [3] show ed that bicycle helmets pr otected their wearers in a sur vey of accide nts in Melbourne, Australia. Rivar a [4] sho wed that bicycle helm ets pr otected their wearers in an exhaustive survey of bicycle accide nts in the Seattle area. Each of these studies also concluded that helmet use had no discernible effect on neck injury. There seemed to be no adver se effect ascr ibable to helme ts. However, ther e ar e limits to wha t broad sur veys can show. The m ajor ity of the incidents discussed in all three of these studies did not notably tax the protective capabilities of the helmets wor n. The results were am ple to prove the value of currently available helmets but they could render no conclusive determination as to whether some helme ts wer e super ior to other s. Even so, one of the charts taken from Rivara' s study showed injury severity plotted against a helmet dam age scale. The plot showed a fairly flat line suggesting that injury was insensitive to increasing helmet damage up to the level of damage the investigators called ' catastrophic' . At this point the injury severity took a sharp upwar d jump. The implication is clearly that there are accidents that exceed the capacities of c urrent headgear and that better bicycle headgear w ould pre vent more deaths and injur ies. Evans and Fr ick [5] have applied an elegant method to select only more ser ious incidents from amo ng those detailed in the US F ederal Gover nment' s Fatal Accident Repor ting System (FA RS). Ander son and Kraus [6] have recently used the same selection technique in a separate study. T he method selects motorcycle accidents involving a passenger and driver in which one or both riders died. The data studied by Anderson and Kraus is not as detailed as that collected by Hur t but it is composed of a much larger number of much m ore serious in cid en ts. H urt noted that ev en inexpensive and unknow n helm ets

25

Becker including at least one antique could provide adequate head protection. On the basis of his study, he concluded that there was no need for m ore stringent standards than the DOT requirem ent. The F ARS data suggest something differ ent. The FAR S data did not identify the headgear used so no direct helmet comparisons were possible. However, Ander son and Kraus showed that motorcycle helmet effectiveness appears to have impr oved steadily over the fourteen year period studied. A lthough they war n that this improvem ent could be due to artifacts and unidentified biases in the data, the authors suggest instead that the increa se in effectiveness ar ose from impr ovements in helmet design over the per iod studied and the replacem ent of inappr opr iate headgear w ith qualified motorcycle helm ets. Better helm ets may indeed pr ovide better protection. The developments in helmets which are traced her e back to Cairns in the 1940' s begin in epidemiology and must be continually reexamined by fresh epidemiological research. The un certainties about test severities and test criteria may never be resolved in any other way. The m ost recent studies suggest clear sailing but there ar e always those who will see rocks ahead and few who w ill insist the boat is already holed and sinking.

26

Becker Bibliography 1. 2. 3. 4. 5. 6. Gurdjian E . S. , Head Injury from Antiquity to the Present with Special Reference to Penetrating Head Wou nds. Springfield, IL : C harles C. Thom as, 1973 Hurt H. H. , Ouellet J.V. , Thom D. R. , Motorcycle Accident Cause Factors and Identification of Counterm easures. Final Repor t Vols 1 and 2, L os Angeles, CA : T raffic Safety Center U SC, 1981 . Williams M . , The P rotective P erform ance of Bicyclists' Helmets in Ac cidents. Accident Analysis & Pr evention 1991:Vol 23. Nos. 2/3. pp. 119-131. Rivara F.P., Thompson D.C. , Thom pson R. S. , Circum stances and Seve rity of Bicycle Injuries. Report, Seattle WA: Har borview M edical Center, 1996. Evans L. and Frick M. C. , Helmet Effectiveness in Preventing Motorcycle Passenger and D river Fa talities. Accident Analysis & Pr evention 1988: Vol 20. pp 363-369. Ander son C. L. and Kraus J. F. , The C hanging Effect of M otorcycle Helm et Use on M ortality: C ompar isons of Driv ers and P asseng ers on th e Sa me M otorc yc le . In: 40th Annual Proceedings Association for the Advancement of Automotive Medicine. Vancouver , BC, 1996:427-442. Cair ns H. , Head Injuries in M otor-C yclists: T he Impor tance of the C rash H elmet. British Medical Journal. 1941: pp. 465-471. Cair ns H. and H olb ou rn H ., Head Injuries in M otor-C yclists with Special Refer ence to C rash H elmets. British M edical Journal. 1943: pp. 591-598. BS 1869:1960 Sp ecification for Protective Helmets for Racing Motorcy clists. London, England: British Standards Institution, 1960. BS 2001:1956 P rotective Helm ets for Motorcyclists. London, England: British Standards Institution, 1956. BS 2095:1958 Industrial Safety Helmets (Light Duty). London, England: British Standards Institution, 1958. BS 2495:1960 Specification for Protective Helmets and Peaks for Racing Car Drivers. London, England: British Standards Institution, 1960. BS 2826:1957 Industrial Safety Helmets (Heavy Duty). London, England: British Standards Institution, 1957. Lombard C. F. , Smith W. A. , Roth H. P. , Rosenfeld S. , Voluntary T olerance of the Hum an to Accelerations of the Head. Journal of Aviation Medicine. 1951: V ol. 22 No. 2 pp. 109-116. Aya R. H. , The L egacy of "P ete" Snell, In: The Book of M otorcycles, Trail Bikes & Scooters/Erik Arctander: Fawcett. 1959 Standards for Racing C rash H elmets. San Fr ancisco, CA : Snell M emor ial Foundation, 1959. Standards for Protective Headgear 1962. San Fr ancisco, CA : Snell M emor ial Foundation, 1962. Standard for Protective Headgear 1970. Sacram ento, CA : Snell M emor ial Foundation, 1970. Standard for Protective Headgear 1975. Sacram ento, CA : Snell M emor ial Foundation, 1975. Standard for Protective Headgear 1980. Sacram ento, CA : Snell M emor ial Foundation, 1980. 1985 Standard for Protective Headgear for U se with Motorcycles and other M otorized Vehicles. Wakefield, RI: Snell Mem orial Foundation, 1985. 1985 Special Applications Stan dard for P rotective Headgear for U se in Co mpetitive Automotive Sports. Wakefield, RI: Snell Mem orial Foundation, 1985. 1990 Standard for Protective Headgear for Use with Motorcycles and other M otorized Vehicles. St. Jame s, NY : Snell M emor ial Foundation, 1990. 1990 Special Applications Standard for Protective Headgear for Use in Compe titive Automotive Sports. St. James, NY : Snell M emor ial Foundation, 1990. 1995 Standard for Protective Headgear for Use w ith Motorcycles an d other Motorized Vehicles. Nor th Highlands, CA : Snell M emor ial Foundation, 1995. 1995 Special Applications Standard for Protective headgear for Use in Competitive Automo tive Sports. No rth Highlands, CA : Snell M emor ial Foundation, 1995. Aya R. R. , He ad P rotection: Safety Goes to the R aces, The M agazine of Standards. September 1966: pp. 256261.

7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

27

Becker 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. Z90. 1-1966 American Standard Specifications for Protective H eadgear for Vehicular Users. New York, NY: Am erican Standards A ssociation, Inc. , 1966. Z90. 1-1971 Am erican National Standard Specifications for Protective Headgear for Vehicular U sers. New Yor k, NY : A merican N ational Standards Institute, Inc. 1972: includes Z90. 1-1973a. 1973. ISO R 1511 Protective Helm ets for Road Users. Switzerland: International Standards O rganization, 1970. ISO DIS 6220-1983 Headforms for Use in the Testing of Protective H elmets. International Standards Or ganization, 1983. Regulation 22 Revision 4 Uniform Provisions Concerning the Approval of P rotective Helmets and of their Visors for Drivers and P assengers of motor Cycles and M opeds. Geneva, Switzerland: United Nations, 1996. National Highwa y Tr affic Safety Administration, M otorcycle Helm ets Proposed M otor Vehicle Safety Standard, Federal Register, M ay 1972, Vol. 37, No. 98, pp. 10079-10085. National Highwa y Tr affic Safety Administration Departm ent of Tr ansportation, Federal Motor Vehicle Safety Standar ds M otorcycle He lmets. Federal Register, August 1973, Vol. 38, No. 160, pp. 22390-22396. National Highway Tr affic Safety Administration Department of Tr ansportation, Federal M otor Vehicle Safety Standar ds M otorcycle He lmets. Federal Register, January 1974, Vol. 39, No. 19, pp. 3554-3555. Hurt H. H. , Thom D . R. , Smith T. A. , Updating the Twenty-Year Old DOT H elmet Sta ndard, In: 40th Annual Proceedings Association for the Advancement of Automotive Medicine. Vancouver , BC, 1996:443-461. BS 4544:1970 Sp ecification for Protective Helmets for Peda l Cyclists. London, England: British Standards Institution, 1970. Standard for Protective Headgear Appendix III Bicycle Helmets 1973. Sacramento, CA : Snell Memor ial Foundation, 1973. 1984 Standard for Protective H eadgear for U se in Bicycling. Wakefield, RI: Snell Mem orial Foundation, 1984. 1990 Standard for Protective Headgear for Use in Bicycling. St. Jam es, NY : Snell M emor ial Foundation, 1990. 1995 Standard for Protective Headgear for Use in Bicycling. St. James, NY : Snell M emor ial Foundation, 1990. Z90. 4-1984 American National Standard for Protective H eadgear - for Bicyclists. New York, NY : A mer ican National Standards Institute, Inc. 1984. ASTM 1447-94 Standard Specification for Protective H eadgear Used in Bicycling. Philadelphia, PA: Am erican Society for Testing and M aterials, 1994. Docum ents Related to Pe tition C P 90-1, Bicycle H elmets. Washington, DC : Consum er Pr oduct Safety Com mission, 1991. Fenner H . A . and James A . F . , JF73 Standard Recommended Te st Method and Performance Standard for Protective H eadgear for Football Players. Hobbs, NM , 1973. ASTM F 717-89 Standard Specification for Football Helmets. Philadelphia, P A: Am erican Society for Testing and M aterials, 1989. Gadd C. W. , Use of a W eighted-Impulse Criterion for Estimating Injury H az ard. In: Conference Proceedings 10th Stapp Car Crash Conference. Holloman A ir F orce Base, NM , 1966, pp. 95-100. FMVSS 208 O ccupant Crash Protection. Washington, DC : N HT SA Unter harnscheidt F.J., Translational versus Rotational Acceleration: A nimal Experiments with Measured Input. Scandinavian Journal Rehabilitative M edicine 4, 1972, pp. 24-26. Collected Do cum ents Related to Swing Aw ay Rig, No rth Highlands, CA : Snell M emor ial Foundation Files Walker, L. B., Harris, E. H. , Pontius U.R. , M ass, Volume, Center of Mass, and M ass Mom ent of Iner tia of Head and Head and N eck of H uman Body. In: Proceedings of Seventeenth Stapp Car Crash Conference, Oklahom a City, OK, 1973. Beier G., Schuck M. , Schuller E., Spann W. , Ew ing C. L. , Becker E. B. , Thom as D. J. , Deter mination of Physical Data of the Head I: C enter of Gravity and Mom ents of Inertia of Human Heads. ONR Report N-00014-75-C-0486. Washington, DC , 1979.

52.

28

Becker 53. Saczalski K.J., States J.D. , Wagar I. J. , Richardson E . Q. , A C ritical Assessment of the Use of Non-Hum an Responding Su rrog ates fo r Safety Syste m Eva luation. In: Proceedings of Twentieth Stapp C ar Crash Conference. Dear born, M I, 1976. JIS T 8133-1982 Japanese Industrial Standard Protective H elmets for Vehicular Users. Tokyo, Japan: Japanese Standards Association, 1982. JIS T 8134-1982 Japanese Industrial Standard Protective Helmets for Bicycle U sers. Tokyo, Japan: Japanese Standards Association, 1982. AS 25112. 9-1990 Australian Standard Methods of Testing Protective Helmets M ethod 9: Determ ination of Resistance to Localized Loading. Sydney, NSW , Australia: Standar ds Australia, 1990. KO VFS 1992:2 G uidelines for Bicycle Helm ets. Swede n: National Swedish Boar d for C onsum er P olicies, 1992. BS 6685:1985 Protective Helm ets for Vehicle Users. London, England: British Standards Institute, 1985. Glaister D. H. , The D evelopm ent and Initial Evaluation of an Oblique-Im pact Te st for P rotective H elmets. Pr esented at: NA TO AG ARD , Cologne, Ger many, 1982.

54. 55. 56. 57. 58. 59.

29